COMPLEX ALGORITHM FOR RECOVERING BINARY IMAGES BASED ON MORPHOLOGICAL METHODS OF GRAPHIC DATAPROCESSING

Abstract

The analysis of modern approaches used in the building and optimization of software and neural network algorithms
for recovery, segmentation and analysis of graphic data is carried out. The priority of morphological methods for processing
binary images in conditions of limited computing resource of the machine analysis system with a large volume of incoming
requests and the necessity for their processing in real time regime is shown. The mathematical model of morphological
processing of binary images based on logical operators of negation, disjunction, conjunction, implication, equivalence,
antiequivalence and difference is proposed, and the relationship between these operators is shown at the level of the
mathematical apparatus. On the basis of the presented mathematical model, typical algorithms for processing binary images
(dilatation and erosion procedures) are constructed. An algorithm for evaluating the effectiveness and optimization of these
procedures has been developed in accordance with the quality indicators of image restoration and the processing time of the
input data, which is based on the classification of a set of queries and the determination of the extrema of the objective
functions.